_samplers

cultivars.engine._core._samplers = <module 'cultivars.engine._core._samplers' from '/home/runner/work/cultivars/cultivars/src/cultivars/engine/_core/_samplers.py'>

single conditional draws with no recursion over data.

The membership test for this module is the same one that keeps the Hamilton filter out of cultivars._core: a primitive may draw from a distribution or vectorize arithmetic over a sample, but the moment a function walks the time axis – a filter pass, a smoother pass – it is an engine and belongs in the internals layer. The stochastic-volatility path draw therefore lives in cultivars._internals._samplers, composing these primitives with the state-space engine; what lives here is everything it needs that has no memory: the Kim-Shephard-Chib mixture constants and indicator draw, and the conjugate covariance draws.

References

Kim, S., Shephard, N., & Chib, S. (1998). Stochastic volatility:

Likelihood inference and comparison with ARCH models. Review of Economic Studies, 65(3), 361-393.

Type:

Sampling primitives